This article explains Apache Pulsar’s fast‑growing adoption, its compute‑storage separation architecture, BookKeeper‑based persistence, multi‑tenant support, flexible subscription models, and fault‑tolerant design, providing a comprehensive overview for developers interested in modern distributed messaging middleware.
This article describes how the QMQ delay‑message service was refactored by separating business logic from storage, adopting Apache BookKeeper for high‑availability, zone‑aware disaster recovery, a configurable DNS resolver, a ZooKeeper‑based task coordinator, and a multi‑level sliding‑time‑bucket scheduler to achieve a stateless, elastic architecture.
This article provides a comprehensive overview of Apache Pulsar, detailing its multi‑tenant architecture, core components such as brokers and BookKeeper, message handling features, producer and consumer configurations, comparison with Kafka, and operational considerations for deployment.
Apache Pulsar is a cloud‑native distributed messaging system that combines stateless brokers, a BookKeeper storage cluster, and ZooKeeper metadata to route messages from producers to consumers across scalable partitioned or non‑partitioned topics, supporting multiple subscription models, producer access modes, routing strategies, batching, acknowledgments, delayed delivery, and deduplication.
This article explains how Apache Pulsar brokers track consumer acknowledgment using cursors, describes persistent and non‑persistent subscription differences, details cursor metadata stored in ZooKeeper and BookKeeper, and outlines optimizations for handling message gaps, including RangeSet improvements and a new LRU‑based storage design.
This article explains how Apache Pulsar brokers track consumer acknowledgments using cursors, store metadata in ZooKeeper and BookKeeper, handle acknowledgment gaps with range sets, and optimize memory usage through LRU and segmented storage to ensure reliable message delivery.
Apache Pulsar is a cloud‑native distributed messaging platform that combines messaging, storage, and lightweight compute, featuring multi‑tenant support, geo‑replication, and high throughput, and this article introduces its core concepts, architecture components such as brokers, BookKeeper, ZooKeeper, and key design features.
This article explains Apache Pulsar’s message storage architecture in BookKeeper, details the ledger and entry lifecycle, describes the multi‑layer caching read path, and outlines BookKeeper’s garbage‑collection process along with practical operational tips for avoiding disk‑heavy scenarios.
This article explains Apache Pulsar’s core architecture—including its message model, producer and consumer behavior, subscription types, storage hierarchy, acknowledgment mechanisms, cursor handling, backlog concepts, retention policies, and storage size calculations—helping readers understand why storage size can exceed backlog size.
This article explains how Apache Pulsar’s layered architecture separates I/O modes, uses BookKeeper for durable storage, and leverages multi‑replica write quorums and hierarchical caches to provide efficient tail‑read, catch‑up read, and long‑term storage options such as S3, Azure and GCS.
